DeepSeek V4 Flash 0731 reap 150b needs about 75.5 GB of memory at IQ3_XXS (a 67.0 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 5 quants, from 0.3 GB to 85.0 GB (MXFP4_MOE).
DeepSeek V4 Flash 0731 reap 150b – 150 billion‑parameter LLM with 132 routed experts
DeepSeek V4 Flash 0731 reap 150b is a 150 billion‑parameter model with 132 routed experts per layer, reduced from 256 by REAP pruning. It is released under the MIT license and supports standard text generation with built‑in thinking. The model is available in several GGUF quantizations.
Good at
- Large‑scale reasoning
- Long‑form text generation
- Built‑in thinking mode
Limits
- No multi‑token‑prediction modules – speculative decoding unavailable
- Requires 8 k context for memory estimates
- No vision or multimodal support
Best for
- Research on large‑parameter LLMs
- Deploying on GPUs with 8–16 GB VRAM
- Testing reasoning capabilities
Who it’s for. Engineers and researchers who need a large‑parameter model with routed experts and built‑in thinking. The model is suitable for experimentation and deployment on systems that support llama.cpp.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 0.3 GB | 0.8 GB | Fits | Get file → | |
| Q2_K | 62.4 GB | 70.4 GB | Too big | Get file → | |
| IQ3_XXS | 67.0 GB | 75.5 GB | Too big | Get file → | |
| Q3_K_M | 70.9 GB | 80.0 GB | Too big | Get file → | |
| MXFP4_MOE | 85.0 GB | 95.8 GB | Too big | Get file → |
For 8 GB VRAM, use the Q2_K quantization (~70 GB file, ~70 GB memory at 8 k context). For 16 GB VRAM, Q3_K_M (~71 GB file, ~80 GB memory) is a good balance. The MXFP4_MOE baseline (~85 GB file) requires the largest memory and is best for high‑end GPUs.
What GPU do you need for DeepSeek V4 Flash 0731 reap 150b?
These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.
- 4 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 6 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 8 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 12 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 16 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 32 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
- 48 GB VRAM: the best quant that fits is the main file (0.3 GB file, about 0.8 GB in use).
Which DeepSeek V4 Flash 0731 reap 150b quant should you choose?
IQ3_XXS (67.0 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
If memory is tight, the smallest file is the first one in the table at 0.3 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about DeepSeek V4 Flash 0731 reap 150b
How much VRAM does DeepSeek V4 Flash 0731 reap 150b need?
About 75.5 GB at IQ3_XXS with an 8K context window. That is the 67.0 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.
Can I run DeepSeek V4 Flash 0731 reap 150b on an 8 GB GPU?
Yes, with the smaller files: the 0.3 GB file needs about 0.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of DeepSeek V4 Flash 0731 reap 150b GGUF?
It depends on the quant: file is 0.3 GB, Q2_K is 62.4 GB, IQ3_XXS is 67.0 GB, Q3_K_M is 70.9 GB, MXFP4_MOE is 85.0 GB.
Where can I download DeepSeek V4 Flash 0731 reap 150b GGUF?
The files are on Hugging Face in the puwaer/DeepSeek-V4-Flash-0731-reap-150b-gguf repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.
What license does DeepSeek V4 Flash 0731 reap 150b use?
Hugging Face lists the license as mit. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.
Can I run DeepSeek V4 Flash 0731 reap 150b on 8 GB VRAM?
The Q2_K quantization requires about 70 GB of VRAM at 8 k context, so 8 GB VRAM is insufficient.
Does this model support vision or multimodal inputs?
No, the model has no vision or multimodal capabilities.
What is the impact of removing multi‑token‑prediction modules?
Speculative decoding is unavailable; standard generation proceeds normally.